Wednesday, December 13, 2006

The Cylons are coming!

Bill Gates, not content with trying to extract a nickel from every link you click, now wants to risk the inevitable nuclear war that we all know follows the proliferation of robots. But I think Bill misses the point entirely when he says this:
Although a few of the robots of tomorrow may resemble the anthropomorphic devices seen in Star Wars, most will look nothing like the humanoid C-3PO. In fact, as mobile peripheral devices become more and more common, it may be increasingly difficult to say exactly what a robot is. Because the new machines will be so specialized and ubiquitous—and look so little like the two-legged automatons of science fiction—we probably will not even call them robots.
Sure, a lot of the new robots will be something like the Roomba vacuum. But there's a perfectly obvious reason why some humanoid robots will be made in large numbers - the world is currently built for humanoids humans. Compatibility matters, something Bill Gates of all people should know really well.

The reality is that robots have already replaced human workers to a huge degree in industry, wherever possible. So the big pool of labour left to be automated is in services - fast food, retail, and the rest. That's the market for automation, something that's already happening to some extent with e-commerce. People might not buy a book from a robot's recommendation, but the robot will probably be able to find a book in a Borders or Chapters quicker, especially when every book has a built-in RFID tag. And I'd wager my customer service experience at McDonalds would get much, much better with robots.

Moreover, if the evolution of the personal computer shows us anything, it's that people prefer a more expensive, but more general-purpose, machine to a cheaper, more specialized one. And when it comes to general-purpose labour, the human form is pretty good, especially for the things we want labour to do. Sure, give it night vision and an extra set of arms if you like. Maybe a tail - for balance. Or coolness.

There's also the issue of cultural expectations, of course. If you've perfected the domestic robot, my suggestion would be to license the image of C3PO and the voice of Anthony Daniels. You'd make a mint from the early adopters who, let's face it, are going to be heavy on the nerd factor.

So assume we nerds get our wish, and cheapish robots proliferate, with a variety of skills available. Where are they going to be used? The answer to that question is pretty simple - anywhere they can be. Here we have the last 30 years of economic history as a guide - robots replacing any workers they can. But that's the crucial question - where will we let them be used?

You'd think that some of the most labour-intensive skills left in our economy (education, health care, the military) would be good candidates, but all three cases are highly regulated and have been captured by the incumbents in the market. As much as the US government could use robo-GIs in large numbers right about now, you only have to look at the institutional hostility the Air Force has displayed to unmanned aerial vehicles to see how well that will fly. And the idea of parents letting their children be taught by a robot tutor doesn't strike me as likely, certainly not in the beginning. Too much Frankenbaggage.

The robot butler/maid is a staple of SF fiction, but there's a pretty good economic rationale for it coming early - its something most middle-class households could use, and if it's got any kind of decent lifespan, the costs could be well less than hiring the human variant. Moreover, it's a relatively free market. I think the more interesting question will be how we would adjust our regulations and laws to accommodate robots? Do we give them driver's licenses? Why not - they're bound to be better drivers than all but the best humans. (An interface with the raft of sensors new cars already have would be incredibly useful.) Besides, how else will they do the grocery shopping?

The other big area where I could see a lot of use for cheap, multi-purpose robots is in agriculture and meatpacking. It's no secret that fruit and vegetable growing in North America is heavily dependent on illegal (cheap) labour, and meatpacking is one of the dirtiest, most dangerous jobs around. Not only would automating those jobs mean fewer miserable jobs in the world, but probably fewer dead people.

(As a side-note, some of the most environmentally friendly, super-efficient forms of agriculture are also the most labour-intensive. Farm robots could dramatically raise food output without sacrificing the environment.)

But robots would have to be really, really cheap for us to get to this point. For general industry, North American workers are already in competition with existing robots, not to mention 12-year old Chinese girls in Shenzhen. Agricultural labour is so cheap it's literally criminal - something that has spread to plenty of other industries across the US, like homebuilding and waitering. I can't imagine why anyone would adopt machines unless a) they were qualitatively better than human workers (plausible to likely) or b) they were substantially cheaper, and easily substituted.

No comments: